I have a matrix Q with 1,480 rows and 740 columns. I have a vector op3 with 740 rows and 1 column. As expected, Q.op3 gives a resulting vector of 1480 elements. However, MathCad, which is normally very picky about having compatible dimensions, is quite happy to multiply Q transpose by op3 and get an answer!! I want to know (a) how this is allowed, and (b) what is it actually doing to get the result? Please see screenshot of MathCad worksheet. Thank you in advance.
Same behaviour for small example